The right answer for the wrong question: consequences of type III error for public health research.

Abstract

This study examined the impact of assessing the causes of interindividual variation within a population when the research question of interest is about causes of differences between populations or time periods. This discrepancy between the research focus and the research question is referred to as a type III error, one that provides the right answer for the wrong question. Homelessness, obesity, and infant mortality were used to illustrate different consequences of type III errors. These different consequences depend on the relationships between the causes of within- and between-group variation. The causes of inter-individual variation and the causes of variation between populations and time periods may be distinct. The problem of examining invariant causes deserves attention.